Brain parenchymal fraction in healthy individuals and in clinical follow-up of multiple sclerosis

Abstract: Background Multiple sclerosis (MS) is an autoimmune disease characterised by inflammatory damage to the central nervous system (CNS). Accumulated CNS injury can be quantified as brain atrophy, definable as a reduction in brain parenchymal fraction (BPF). BPF correlate with disability in MS and is used routinely as an endpoint in clinical trials. In 2009/2010, a new MS clinical care program, that includes follow-up of BPF, was introduced at Umeå University Hospital (NUS). Levels of neurofilament light polypetide (NFL) and glial fibrillary acidic protein (GFAP) in cerebrospinal fluid (CSF) are markers of axonal and astrocytic injury, respectively, and also potential surrogate biomarkers for BPF decline. The goals of this thesis were to establish age-adjusted values of BPF in healthy individuals and to relate these to the BPF values from individuals with MS as well as to the levels of NFL and GFAP in CSF. Another goal was to investigate if expanded disability status scale (EDSS)-worsening could be predicted in a clinical MS cohort and if BPF measurements could contribute to such predictions. Methods A group of 111 healthy individuals volunteered to participate in the studies. A total of 106 of these underwent MRI with BPF measurements, 53 underwent lumbar puncture (LP) with measurement of NFL and GFAP and 48 underwent both MRI and LP. Three different automatic and one manual method were utilised to determine BPF. A literature search on BPF in healthy individuals was performed for the purpose of a systematic review. For studying disability progression in MS, all individuals with MS followed at NUS and included in the Swedish MS registry were included if they had matched data on BPF, EDSS and lesion load as part of clinical follow-up (n=278). Results BPF as well as NFL and GFAP levels in CSF were all associated with age. NFL was associated with BPF and GFAP, but only the association with GFAP was retained when adjusting for age. Significant differences were found between different methods for BPF determination. In the MS population, BPF was associated with EDSS. Only progressive disease course could predict EDSS worsening. Conclusion The data on BPF and levels of NFL and GFAP in CSF of healthy individuals can aid in the interpretation of these variables in the setting of MS. Knowledge on differences in BPF data from different methods for BPF determination can be useful in comparing data across studies, but also highlights the need for a commonly accepted gold standard. The correlation between GFAP and NFL levels in CSF may indicate an association between glial and axonal turnover that is independent of the aging effect on the brain. However, the low number of volunteers for LP precluded clear conclusions. An association between BPF and EDSS was seen in the MS group. The ability to predict EDSS worsening in the clinical MS cohort was limited.